Créditos
6
Tipos
Optativa
Requisitos
Esta asignatura no tiene requisitos
, pero tiene capacidades previas
Departamento
CS
El temario revisa algunos fundamentos y profundiza en diversas técnicas modernas de aprendizaje no lineal, desde redes neuronales modernas hasta métodos avanzados de aprendizaje basado en kernel y los últimos avances en métodos de conjunto. También busca proporcionar una visión unificada del área y sus posibles perspectivas futuras.
Profesorado
Responsable
- Luis Antonio Belanche Muñoz ( belanche@cs.upc.edu )
Otros
- Jamie Arjona Martínez ( jamie.arjona@upc.edu )
Horas semanales
Teoría
3.2
Problemas
0
Laboratorio
1
Aprendizaje dirigido
0
Aprendizaje autónomo
7.38
Competencias
Uso solvente de los recursos de información
Lengua extranjera
Básicas
Genéricas
Específicas
Objetivos
-
Métodos avanzados de aprendizaje automático
Competencias relacionadas: CB10, CB6, CT4, CT5, CE10, CE8, CE9, -
Estadística bayesiana
Competencias relacionadas: CB7, CT4, CT5, CE10, CE5, CE8, -
Optimización de redes neuronales y máquinas de vectores soporte
Competencias relacionadas: CB10, CB6, CB7, CT4, CT5, CE10, CE11, CE3, CE5, CE8, CE9, CG2, -
Modelos lineales y modelos lineales generalizados de regresión no paramétricos
Competencias relacionadas: CB10, CT5, CE10, CE5, -
Limpieza de datos
Competencias relacionadas: CB6, CT4, CE11, CE3, CE8, CE9, CG2,
Contenidos
-
Theoretical refresher of machine learning. Introduction to Bayesian machine learning.
Introduction to Bayesian thinking for machine learning. Learning by solving a regularized problem. Illustrative example. -
Learning in functional spaces
Reproducing kernel Hilbert spaces. The representer theorem. Example 1: Kernel ridge regression. Example 2: The Perceptron and the kernel Perceptron. -
Fundamental kernel functions in R^d.
Description and demonstration of fundamental kernel functions in R^d. Polynomial and Gaussian kernels. General properties of kernel functions. -
The support vector machine for classification, regression and novelty detection
The support vector machine (SVM) is the flagship in kernel methods. Its versions for classification, regression and novelty detection are fully explained and demonstrated. -
Kernel functions for diferent data types
Some kernel functions for different data types are presented and demonstrated, such as text, trees, graphs, categorical variables, and many others. -
Other kernel-based learning algorithms
Additional kernel-based learning methods are explained, such as kernel PCA and kernel FDA. These are illustrated in several application examples. -
Introduction to deep neural networks. Autoencoders and Variational Autoencoders.
Introduction to deep neural networks: reminder of fundamental neural network theory and optimization, qualitative description, loss functions, activation functions, regularization and best practices.
Autoencoders and Variational Autoencoders. -
Special networks: (New) Hopfield neural networks and KANs
Special networks: (New) Hopfield neural networks and KANs. -
Ensemble methods: baggers, boosters and stackers
This activity cover the basic and modern developments in ensemble methods, including baggers, boosters and stackers. -
Advanced and hybrid techniques in deep networks and kernel methods
Other methods are briefly introduced, such as the RVM and GPs. Nyström acceleration and Random Fourier features. Deep kernel learning and maybe others.
Actividades
Actividad Acto evaluativo
Theoretical lectures
Objetivos: 1 2 4 3
Contenidos:
- 1 . Theoretical refresher of machine learning. Introduction to Bayesian machine learning.
- 2 . Learning in functional spaces
- 3 . Fundamental kernel functions in R^d.
- 4 . The support vector machine for classification, regression and novelty detection
- 5 . Kernel functions for diferent data types
- 6 . Other kernel-based learning algorithms
- 7 . Introduction to deep neural networks. Autoencoders and Variational Autoencoders.
- 8 . Special networks: (New) Hopfield neural networks and KANs
- 10 . Advanced and hybrid techniques in deep networks and kernel methods
Teoría
40h
Problemas
0h
Laboratorio
0h
Aprendizaje dirigido
0h
Aprendizaje autónomo
25h
Practice lectures
Objetivos: 1 3 5
Contenidos:
- 1 . Theoretical refresher of machine learning. Introduction to Bayesian machine learning.
- 3 . Fundamental kernel functions in R^d.
- 4 . The support vector machine for classification, regression and novelty detection
- 5 . Kernel functions for diferent data types
- 6 . Other kernel-based learning algorithms
- 7 . Introduction to deep neural networks. Autoencoders and Variational Autoencoders.
- 8 . Special networks: (New) Hopfield neural networks and KANs
Teoría
0h
Problemas
0h
Laboratorio
16h
Aprendizaje dirigido
0h
Aprendizaje autónomo
16h
Term project
Objetivos: 1 2 4 3 5
Contenidos:
- 1 . Theoretical refresher of machine learning. Introduction to Bayesian machine learning.
- 2 . Learning in functional spaces
- 3 . Fundamental kernel functions in R^d.
- 4 . The support vector machine for classification, regression and novelty detection
- 5 . Kernel functions for diferent data types
- 6 . Other kernel-based learning algorithms
- 7 . Introduction to deep neural networks. Autoencoders and Variational Autoencoders.
- 8 . Special networks: (New) Hopfield neural networks and KANs
Teoría
0h
Problemas
0h
Laboratorio
0h
Aprendizaje dirigido
0h
Aprendizaje autónomo
32h
Metodología docente
El curso profundiza en los paradigmas de aprendizaje automático más importantes con una base sólida en probabilidad, estadística y matemáticas. La teoría se introduce en clases magistrales donde el profesor expone los conceptos. Estos conceptos se ponen en práctica en las clases de laboratorio, en las que el alumno aprende a desarrollar soluciones de aprendizaje automático a problemas reales de cierta complejidad.Los estudiantes tienen que trabajar y entregar un proyecto al final del curso.
Método de evaluación
La asignatura se califica de la siguiente manera:F = Nota del examen final
P1, P2, P3 = Nota de los trabajos prácticos (1, 2 y 3)
Nota final = 25% F + 25% P1 + 25% P2 + 25% P3
Bibliografía
Básico
-
Pattern recognition and machine learning
- Bishop, Christopher M,
Springer,
cop. 2006.
ISBN: 0387310738
https://discovery.upc.edu/discovery/fulldisplay?docid=alma991003157379706711&context=L&vid=34CSUC_UPC:VU1&lang=ca -
Kernel methods for pattern analysis
- Shawe-Taylor, John; Cristianini, Nello,
Cambridge University Press,
2004.
ISBN: 0521813972
https://discovery.upc.edu/discovery/fulldisplay?docid=alma991002747459706711&context=L&vid=34CSUC_UPC:VU1&lang=ca -
Neural networks and deep learning : a textbook
- Aggarwal, Charu C,
Springer,
2023.
ISBN: 9783031296420
https://ebookcentral-proquest-com.recursos.biblioteca.upc.edu/lib/upcatalunya-ebooks/detail.action?pq-origsite=primo&docID=30620507 -
Deep learning : foundations and concepts
- Bishop, Christopher M; Bishop, Hugh,
Springer,
[2024].
ISBN: 9783031454677
https://ebookcentral-proquest-com.recursos.biblioteca.upc.edu/lib/upcatalunya-ebooks/detail.action?pq-origsite=primo&docID=30853138 -
Ensemble Methods: Foundations and Algorithms
- Zhou. Zhi-Hua,
CRC Press,
2025.
ISBN: 9781003587774
https://www-taylorfrancis-com.recursos.biblioteca.upc.edu/books/mono/10.1201/9781003587774/ensemble-methods-zhi-hua-zhou